SGLang turns Qwen3.8-27B into a decision model that beats Pokémon FireRed at sub-100ms

zhaoran_wang · x · 2026-09-30

The SGLang team turned Qwen3.8-27B into a multimodal decision model that beat Pokémon FireRed's elite four and champion with sub-100ms decisions from live game state. SGLang now offers a native /v1/decisions endpoint to use LLMs/VLMs as classification and scoring models, plus /v1/systemone for Jev-like open models with the TypeSafe SDK.

Original post →

More from Infra

Infra channel →